PROJECT SUMMARY/ABSTRACT The ability to predict the length of time from disease onset to major disease outcomes in individual patients with Alzheimer's disease (AD) has important implications for patient care, the development of interventions and public health. The major aim of the Predictors Study is to further the understanding of AD progression in order to develop predictor algorithms to address this issue. Over the past funding periods, we have followed two clinic-based cohorts of AD patients recruited from three major medical centers, and have made major progress in characterizing the natural history of AD and identifying predictors of disease course. While the Predictors study has had a major impact on our understanding of AD and its progression, the patient cohorts are clinic- based and ethnically homogenous, and the true date of disease onset was unknown. We therefore have assembled and initiated follow-up a new, well-characterized, population-based cohort of ethnically diverse elders with AD. Many of these individuals were followed from a point prior to the onset of AD, so the onset date of clinical AD is known. Another portion are at-risk for AD, allowing us to track the disease process from it preclinical state. We propose to continue intensive follow-up of this cohort in order to validate our previous Predictors study findings in this population-based cohort and to implement new research questions based on novel predictor and outcome variables. We will use linkage to Medicare and Medicaid data to understand correlates of the economic impact of AD in this multiethnic community cohort, focusing on costs associated with transition to dementia, medication utilization, the relation of dementia status to the cost of comorbid conditions, end of life costs, and lifetime economic burden based on predicted disability-free and disabled survival. We will also continue to refine a unique predictive approach that uses longitudinal Grade of Membership (L-GoM) modeling to accurately summarize the AD process. We will validate and apply an updated L-GoM model to external datasets, extend it to include the pre-dementia phase, explore the genetic correlates of heterogeneity of disease course, incorporate AD biomarkers, determine the extent to the model refines analysis of AD treatment effects by applying it to data from recent AD clinical trials, and develop and support software for calculation of the L-GoM model predictions in individual patient data.